Lipoprotein(a) and Cardiovascular Disease Prevention across Diverse Populations

Lipoprotein(a) [Lp(a)] is a low-density lipoprotein(LDL)-like lipid fraction consisting of two subunits: a single apolipoprotein a [Apo(a)] molecule, which is a lipid-rich fraction produced by the liver; and apolipoprotein B-100 (ApoB100), which has an LDL receptor and is also a constituent of very low density lipoprotein(VLDL), intermediate density lipoprotein(IDL), and LDL [1]. Apo(a) consists of 11 different kringle domains, which are subdomains of the protein with a structure that resembles the folded Danish pastry of the same name. While most kringle domains are expressed exactly once in the final Apo(a) quaternary structure, the primary constituent by mass is the kringle IV type 2 (KIV-2) domain, which can be repeated in the Apo(a) molecule from fewer than 13 times to more than 50 times [2]. The expression profile of the KIV-2 domain is predominantly determined genetically in an autosomal dominant fashion, and there is an inverse relationship between the number of KIV-2 repeats and the concentration of Lp(a) in serum [3].

Considerable overlap exists between the structure of the apo(a) molecule and those of plasminogen and tissue plasminogen activator. The kringle IV type 10 has a lysine binding site similar to plasminogen but the cleavage site has a mutation that prevents proteolysis. Lp(a) thus can compete with plasminogen for the binding site on fibrin but since it lacks the proteolytic subdomain it functionally inhibits fibrinolysis [4].

Lp(a) has also been identified as an acute-phase reactant after tissue injury. Lp(a) was positively correlated with IL-6, C-reactive protein, and alpha-1 antitrypsin after acute myocardial infarcts [5, 6]. Apo(a) may also act as an immune suppressor by decreasing recruitment of neutrophils [7].

Lp(a) is a preferred carrier for oxidized phospholipids (OxPL), which are demonstrated to be immunogenic, proinflammatory, and proatherogenic [8]. In a prospective cohort study of 1400 healthy adults followed for 15 years, highest tertile OxPL was associated with hazard ratios of 2.4 for cardiovascular disease (CVD) and 3.6 for stroke compared to the lowest tertile [9]. Lp(a) is believed to exert part of its proatherogenic effects via its carriage of OxPL. Lp(a) is also believed to increase the propensity to form coronary calcification. Korean men and women in the highest Lp(a) quartile were found to have odds ratios of 1.2 and 1.6, respectively, for coronary artery calcium score greater than 0, even after adjusting for LDL and metabolic syndrome [10].

While the expression of Lp(a) is largely determined by genetic inheritance, behavioral factors such as diet and exercise can affect Lp(a) levels to a modest degree. For example, in a study of 37 healthy women fed two low-fat diets, one low in vegetable content and the other high in vegetable content, they experienced a 7% and 9% increase in Lp(a), respectively [11]. Sustained physical exercise can increase Lp(a) levels by 10% for moderate exercise and up to over 100% for high-endurance athletes [12‐15]. The rise in Lp(a) during exercise is likely as an acute-phase reactant in response to soft tissue damage in the musculature and vasculature during repeated movements. Two studies found no significant relationship between exercise and Lp(a) but did note a non-significant trend of higher mean Lp(a) in the active group [16, 17]. Many studies, however, find no relationship between exercise and Lp(a) levels, and the benefits of exercise on cardiovascular health are well established [18‐25].

This article reviews the up-to-date literature related to Lp(a) and its role in cardiovascular disease, ethnic disparities in the levels and pathogenicity of this lipid fraction, clinical monitoring guidelines, and treatment modalities (Table 1). This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.

Lp(a) is believed to be a highly atherogenic lipid fraction with numerous studies showing a strong association between elevated Lp(a) and coronary artery disease, stroke, and aortic stenosis, which will be explored in the following section.

Studies on Lp(a) have leveraged longitudinal data the Copenhagen City Heart Study (CCHS), the Copenhagen General Population Study (CGPS), and the Copenhagen Ischemic Heart Disease Study (CIDHS). All studies contained genetic as well as outcomes data. These studies employ Mendelian randomization, an epidemiological method that uses the independent assortment of risk alleles as a randomization event to estimate the association between a genetic profile and a clinical outcome of interest [26]. Mendelian randomization studies are not able to determine a causal relationship between a gene and a clinical outcome in the same way that randomized controlled trials can, but they can be stronger than normal observational studies at determining causation if the risk allele that is studied is not confounded by linkage disequilibrium or population stratification [27]. In the aforementioned Danish studies, the population in question was ethnically homogenous (i.e., no population stratification effects) and LPA, the gene under study, is not known to be in linkage disequilibrium with any other genes, though this possibility can rarely be excluded.

Persons of African and South Asian descent have higher mean Lp(a) levels [43‐45]. In African Americans, Lp(a) displays a bell-shaped frequency distribution around the median and mean [46]. By contrast, the distribution of Lp(a) concentrations in Caucasians is skewed, with the mean value exceeding the median due to a small number of individuals with very high Lp(a) levels [3, 47]. In the Charleston Heart Study cohort, elderly African-American men demonstrated a more skewed distribution of Lp(a) concentration frequencies, resembling that of the white cohort participants [48]. Thus it may be the case that Lp(a) decreases as African Americans age or that those with elevated Lp(a) are less likely to survive to the 7th, 8th, or 9th generation.

Even at young ages, African-American children have Lp(a) levels that are 1.7 times that of white children [49]. While white children with parental history of MI were more likely to have Lp(a) greater than 25 mg/dl compared to those without parental MI, African-American children did not display a similar pattern, suggesting a different risk profile associated with elevated Lp(a) in African Americans [49]. Indeed, the Multiethnic Study of Atherosclerosis demonstrated a significant relationship between Lp(a) and carotid atherosclerosis for whites, but was equivocal for Hispanics and not significant for African Americans [50].

African Americans have been demonstrated to have favorable total cholesterol and triglyceride profiles compared to Caucasians despite Lp(a) levels that were 2.2 times greater than levels seen in Caucasians [45]. When comparing overweight and obese prediabetic African-American and Caucasian women, African-American women were found to have higher HDL and ApoA1 and lower VLDL but they also had higher CVD mortality than their Caucasian counterparts [51].

The contribution of the apo(a) allele frequencies to Lp(a) levels varies by ethnicity. Ghanaians with the same apo(a) allele polymorphs as Germans still had up to a three-fold increased Lp(a) serum concentration than their German counterparts, and the difference in expression was especially pronounced in patients who had high molecular weight polymorphs [52]. The Coronary Artery Risk Development in Young Adults (CARDIA) trial, a prospective longitudinal epidemiological study of African-American and Caucasian subjects, demonstrated that at very high molecular weights of apo(a), both groups had low Lp(a) levels. By contrast, there was significant difference in Lp(a) in African Americans and Caucasians who had intermediate size isoforms, with African Americans expressing a wider distribution of Lp(a) serum concentrations than Caucasians [53]. The SNP rs9457951 was strongly associated with variation in Lp(a) level in African Americans in the Jackson Heart Study and was confirmed to be influential in the Dallas Heart Study. Of note, this difference was not demonstrated to have an association with increased cardiovascular mortality [54].

In a post hoc analysis of the African Americans and Caucasians in the Atherosclerosis Risk in Communities (ARIC) study, it was demonstrated that African Americans had a wider interindividual distribution of Lp(a) concentrations compared to Caucasians. For African Americans, a race-specific 1-SD–greater log-transformed Lp(a) was associated with a hazard ratio of 1.13 for incident CVD, 1.11 for incident CHD, and 1.21 for ischemic strokes. For Caucasians, the respective hazard ratios were 1.09, 1.10, and 1.07, with ischemic strokes not reaching statistical significance [55].

A retrospective cohort study in East and Southeast Asian populations residing in California found that Indian Americans had higher Lp(a) levels than non-Hispanic whites (36 vs. 29 nmol/l) and that Chinese Americans had the lowest average Lp(a) level (22 nmol/l). There was a trend toward greater burden of ischemic heart disease in subjects with the highest quintile of Lp(a) compared to the lowest quintile, but this relationship was not significant [43].

Unlike HDL, LDL, and triglycerides, Lp(a) is not currently routinely measured in primary care settings. The American College of Cardiology/American Heart Association blood cholesterol guidelines from 2018 consider Lp(a) ≥ 50 mg/dl or ≥ 125 nmol/l to be a risk-enhancing factor and recommend measuring Lp(a) in patients who have a family history of premature atherosclerotic cardiovascular disease (ASCVD) or a personal history of ASCVD not explained by major risk factors [56]. It is recommended to screen Lp(a) levels in women with a history of hypercholesterolemia, but there was minimal improvement in cardiovascular mortality and morbidity risk prediction in adult women with total cholesterol less than 220 mg/dl [57].

One of the main barriers to studying and treating Lp(a) is that measurement is not standardized. Results may vary depending on the specific assay used due to the high degree of polymorphism of Apo(a). Using a monoclonal antibody against the highly variable KIV-2 epitope yields an overestimation of molar concentration for larger Apo(a) molecules and underestimates molar concentration in subjects with smaller Apo(a) phenotypes [58]. A Lp(a) standardization group supported by the World Health Organization and the International Federation of Clinical Chemistry was established to evaluate the extant assays and make recommendations regarding clinical assessment of Lp(a). The group measured Lp(a) serum concentrations using six different assays, providing Lp(a) in mg/dl (Denka Seiken, Abbott Quantia, Beckman, Diasys 21FS, and Siemens N Latex) or in nmol/l (Roche TinaQuant, Diasys 21 FS). There was significant variation between assays at the higher and lower ends of Lp(a) concentration as well as significant intra-assay variability, suggesting that more work needs to be done to standardize Lp(a) assessment across varying labs [59].

Given that Apo(a) isoforms have highly variable molecular weights depending on the number of KIV-2 repeats, it has been suggested that clinicians discontinue using mg/dl units in favor of nmol/l [60]. This would have the added benefit of encouraging further standardization of assays and would align with prevailing evidence that molar concentration is more predictive than volume measurements.

As noted, until recently, there were no drug candidates to target Lp(a) reduction. Statins have not been demonstrated to lower Lp(a) levels. On the contrary, administration of atorvastatin and pravastatin into hepatocytes increased LPA mRNA by 1.5-fold in addition to increasing synthesis of apo(a) [61] (Table 2).

Plasma lipid apheresis is demonstrated to be quite effective at reducing Lp(a), but it also reduces a broader array of lipids and lipoproteins [62, 63]. A German cohort study demonstrated that plasma lipid apheresis reduced MACE by 78% in the year after initiation in assessing patients with combined elevated triglycerides, LDL, and Lp(a) [64]. The German cohort recommends lipid apheresis in patients who have isolated Lp(a) elevations greater than 60 mg/dl absent other lipid derangements. In the UK, Lp(a) can be considered as an additional risk factor to be considered in the initiation of lipid apheresis to treat elevated LDL-C [62]. Lipid apheresis is limited in its practicability due to the significant side-effect profile of fatigue, hypotension, and neurohormonal dysregulation, and thus is only recommended in the treatment of familial hypercholesterolemia in the US, Japan, Australia, and Spain [62].

Niacin has long been known to reduce Lp(a) levels in addition to increasing HDL and reducing LDL and triglycerides. The effect of niacin on cardiovascular outcomes has been called into question after large randomized controlled trials failed to show any benefit. The Heart protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE) trial was a multicenter randomized controlled trial of statin therapy plus either 2 g of extended-release niacin plus 40 mg laropiprant versus placebo in patients with vascular disease. While patients in the treatment arm had higher HDL and lower LDL, they did not experience a significant improvement in nonfatal myocardial infarction, death from coronary causes, stroke, or arterial revascularization after a median follow-up of 3.9 years [65]. Similar lipid profile changes would have been expected to yield a 2% reduction in MACE overall and a 6% reduction in MACE for the highest quintile of Lp(a) [66]. Similarly, the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health Outcomes (AIM-HIGH) trial tested whether extended-release niacin added to intensive statin therapy, as compared with statin therapy alone, would reduce the risk of cardiovascular events in patients with established atherosclerotic cardiovascular disease. The trial demonstrated that while niacin did increase HDL by 7 mg/dl and lower LDL by 12 mg/dl and triglycerides by 42 mg/dl, it did not have a significant impact on the primary composite endpoint of death from coronary heart disease, nonfatal myocardial infarction, ischemic stroke, hospitalization for an acute coronary syndrome, or symptom-driven coronary or cerebral revascularization [67]. The AIM-HIGH trial also showed high rates of adverse side effects due to niacin, including the typical flushing, itching, and raised blood glucose in addition to more serious effects of increased gastrointestinal symptoms and infections [68]. These studies combined have led practitioners to cease prescribing niacin to patients for primary cardiovascular disease prevention.

Proprotein convertase subtilisin/kexin 9 (PCSK9) inhibitors have been demonstrated to reduce Lp(a) levels. A pre-specified analysis of the ODYSSEY outcomes trial of alirocumab in 18,924 patients 1–12 months after acute coronary syndromes showed that PCSK9 inhibition resulted in a 5 mg/dl reduction in Lp(a), a 51.1 mg/dl reduction in LDL-C, and a 15% reduction in MACE [69]. After controlling for the reduction in LDL-C, reducing Lp(a) by 1 mg/dl was associated with a reduction in the hazard ratio for MACE to 0.994 (95% CI 0.990–0.999, p = 0.0081). This suggests that Lp(a) could serve as an independent biomarker for cardiovascular risk. The FOURIER (Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk) trial measured Lp(a) levels in 25,096 patients for a median follow-up of 2.2 years and found that subjects with fourth quartile Lp(a) had a hazard ratio of 1.22 for coronary heart disease death, myocardial infarction, and urgent revascularization compared to those in the first quartile. Evolocumab was compared to placebo in the FOURIER trial, and it was found to reduce coronary heart disease death, myocardial infarction, or urgent revascularization by 23% in patients who had higher than median Lp(a) at baseline and by 7% in patients who had less than median Lp(a) at baseline. The absolute reduction in the composite endpoint was 2.49% and the number needed to treat was 40 over 3 years [70].

Antisense oligonucleotides (ASO) are sequences targeted to reduce expression of specific RNA molecules. One such agent is Mipomersen, an ASO targeted to B-100 that has been demonstrated to reduce Lp(a) levels but is not currently indicated for use outside of LDL reduction [71]. Akcea Pharmaceuticals (formerly Ionis Pharmaceuticals) has developed Apo(a)LRx, an ASO targeted to hepatocytes via conjugation with a triantennary N-acetylgalactosamine (GalNAc3) moiety, a high-affinity ligand for the asialoglycoprotein receptor on the surface of hepatocytes, to reduce hepatic synthesis of Lp(a). A randomized double-blinded controlled phase I trial of efficacy demonstrated that a single dose of the ASO did not result in reductions in Lp(a) but six doses reduced Lp(a) by 39–77% in a dose-dependent manner [72]. The phase 2 trial in 286 patients with cardiovascular disease demonstrated that Apo(a)LRx reduced Lp(a) in a dose-dependent manner, with the highest dose (20 mg every week) reducing Lp(a) by 80% over 27 weeks. The agent appears to be well tolerated, with no significant derangements in liver or renal measures and no reductions in platelet counts. The two deaths in the study were from a car accident and suicide from previously diagnosed depression.